{
    "epoch": 1.0,
    "eval_logits/chosen": -3.5957722663879395,
    "eval_logits/rejected": -3.6472697257995605,
    "eval_logps/chosen": -297.0597839355469,
    "eval_logps/rejected": -260.8578186035156,
    "eval_loss": 0.6086059808731079,
    "eval_rewards/accuracies": 0.697604775428772,
    "eval_rewards/chosen": 0.13394081592559814,
    "eval_rewards/margins": 0.22443543374538422,
    "eval_rewards/rejected": -0.09049463272094727,
    "eval_runtime": 219.8901,
    "eval_samples": 2000,
    "eval_samples_per_second": 9.095,
    "eval_steps_per_second": 0.759,
    "train_loss": 0.6369124913807982,
    "train_runtime": 9387.3825,
    "train_samples": 61966,
    "train_samples_per_second": 6.601,
    "train_steps_per_second": 0.034
}